1. Introduction
The issue of the correct way to conceptualize organismality is typically investigated via examination of problematic cases. Such cases are usually plants, microbes, or other organisms, such as Portuguese man-of-war, only distantly related to homo sapiens. By contrast, mammals, or even all higher metazoans, have been viewed as relatively unproblematic regarding the question of counting organisms. For instance, Godfrey-Smith uses homo sapiens as his illustrative example of a “paradigmatic” individual organism (Reference Godfrey-Smith2009, 95). More recently, this position is challenged by literature on symbiotic gut flora in mammals and the so-called holobiont theory, according to which the genuine organism is what is usually taken to be the whole organism plus its symbiotic microbial communities (Dupré and O’Malley Reference Dupré and O’Malley2009; Gilbert, Sapp, and Tauber Reference Gilbert, Sapp and Tauber2012).
In what follows, I demonstrate that the case of placental pregnancy diversifies examples of “problems of individuality”Footnote 1 in mammals and expands them beyond the microscale. In addition to highlighting this novel, “close to home” case in which counting organisms is not straightforward, I contribute to a biologically informed philosophy of pregnancy.Footnote 2
Section 2 briefly introduces the general problem of counting organisms and examines its relation to Elselijn Kingma’s metaphysics of pregnancy project. Section 3 presents and applies Clarke’s (Reference Clarke2013) evolutionary account of the organism to pregnancy. I demonstrate that this approach yields the clear verdict that there are two organisms, on the grounds that both the gestator and fetus, but not the gestator-fetus, can be targets of a natural selection process in a “Darwinian population” (Godfrey-Smith Reference Godfrey-Smith2009).Footnote 3 They are distinct targets due, principally, to processes of bottlenecking, germ-soma sequestration, and sexual recombination. Section 4 provides a contrasting assessment by applying to pregnancy Pradeu’s (Reference Pradeu2012) immunological theory. I show that the immunological approach yields the answer that there is one organism, because of pervasive interactions across the placenta.Footnote 4
Section 5 concludes that the different perspectives offer different answers to the question, How many organisms are present during a placental pregnancy? Hence, pregnancy is a novel example of the problem of counting organisms; novel because it is mammalian but, unlike the holobiont, is at the scale of “macrobes” (Dupré and O’Malley Reference Dupré and O’Malley2009, 3) of the same species. Additionally, the question of whether holobionts are genuine organisms remains contested (Skillings Reference Skillings2016; Bourrat and Griffiths Reference Bourrat and Griffiths2018). Section 5 also briefly examines which account of the organism is most relevant to a philosophy of pregnancy.
2. Theories of the Organism and Metaphysics of Pregnancy
The issue of how to individuate organisms is central and fundamental to the philosophy of biology. Clarke (Reference Clarke2010) presents only a subset of many proposed criteria for counting organisms, demonstrating that they draw different conclusions about various problematic cases such as Portuguese man-of-war colonial “individuals,” aspen, aphids with mutualistic bacteria, and eusocial insect colonies. This is a problem because in various biological disciplines, from evolutionary to ecological theory, it is crucial to be able to count organisms. Discussion of this subject typically proceeds via examination of the biology of one or another problematic species. One of this article’s aims is to show that mammalian placental pregnancy is an underconsidered example of a problematic case when it comes to counting organisms. As evidence of this lack of consideration of pregnancy in the philosophy of biology literature on individuating organisms, take Clarke’s introduction of her “unproblematic” puppy. “They share with humans all the features that have been historically taken to be reliable hallmarks of individuality. Any definition that excludes this example will be strongly counterintuitive. We’ll assume it’s a ‘normal’ healthy puppy—it doesn’t have any twins, transplanted organs, or tumors, although each of these suppositions would be interesting in their own right” (321).Footnote 5 Regardless of whether pregnancy counts as abnormal, a pregnant dog would, I contend, be an equally interesting supposition. Take too Godfrey-Smith’s comment on “dramatic case(s)” (Reference Godfrey-Smith2009, 78) with complex life histories. “Protists, primitive plants, fungi, and invertebrate organisms often go through elaborate sequences of states, increasing or decreasing their number of chromosomes, fusing or fragmenting, occupying very different environments” (78). Again, I contend, while perhaps not being quite as awkward to tackle as some of the above phenomena, placental pregnancy is a similarly dramatic but common phenomenon with which many of us have firsthand experience.
In addition to a novel application of theories of organismality, my other aim is to strengthen the foundation of a biologically informed metaphysics of pregnancy. Kingma (Reference Kingma2018, Reference Kingma2019) argues against a mere “container” model of mammalian pregnancy that, she claims, has floated free from consideration of the actual biology of pregnancy. Instead, she argues for a part-whole relation between gestator and fetus. Relevant to this article, her argument is based heavily on the claim that both gestator and fetus fail to meet various criteria of organismality proposed in the literature but that the gestator-fetus whole does satisfy some of these criteria.
This article contributes to the metaphysics of pregnancy in two ways. First, it engages more fully with the literature on organisms than Kingma’s initial exploration. When it comes to section 3’s evolutionary approach, Kingma’s (Reference Kingma2019) analysis does not focus on the mechanisms that are most important for individuating organisms. She neglects “bottlenecking,” germ-soma differentiation, and sexual recombination, which form the core of my analysis.
My other contribution to the metaphysics of pregnancy is to separate Kingma’s mereological thesis from what I take to be the related but distinct “counting question”: How many organisms are present during placental pregnancy? It might appear that the two issues are inextricably tied together, in that an answer one implies that the part-whole thesis is correct and an answer two implies that the part-whole thesis is false. In fact, matters are not so straightforward. To claim that a two-organism result implies the falsify of the part-whole claim is to prematurely assume that one organism cannot be part of another organism. The one-organism answer does not immediately vindicate the part-whole claim because we cannot rule out a priori that the fetus is not part of the gestator but does not yet constitute an organism. This is not to say that the counting question is irrelevant to the mereological question, only that the relationship between them is complex and beyond the scope of this article. However, I lay the groundwork here for that more detailed examination by clearly separating out and answering the counting question.
Before moving to evolutionary organismality, a note on terminology. Much of the literature labels the concept I refer to as organismality as “biological individuality.” Indeed, Clarke’s papers use “biological individual” in their titles, while Pradeu predominantly uses “organism.” That said, Clarke explicitly “use(s) the terms … interchangeably” (Reference Clarke2013, n. 1) and switches between them in her papers, while Pradeu often uses the term “individual” when picking out a single organism (Reference Pradeu2012, 227). Yet such switching is somewhat problematic because various authors argue that the two terms are not coextensive (Pradeu Reference Pradeu2016a; Wilson and Barker Reference Wilson, Barker and Zalta2018). One reason for this is that Hull’s (Reference Hull1978) influential “species as individuals” position does not claim that species are organisms, making “biological individual” a more inclusive term than “organism.” It is beyond the scope of this article to adjudicate on the relation between the two terms, but I use “organism” throughout for consistency.
3. Pregnancy and an Evolutionary Approach to Organismality
Recently, Clarke (Reference Clarke2013), among others, developed an evolutionary approach to the question, What is an organism? I think that other evolutionary accounts, such as Godfrey-Smith’s (Reference Godfrey-Smith2009), yield the same judgment regarding pregnancy, but I focus on Clarke in this article. This is because she explicitly claims that her approach unifies various other ones and because it is unambiguously evolutionary in its perspective.
Clarke’s central claim is that an entity is an organism if and only if it can be the target of a process of natural selection. She reaches this definition via rejecting the necessity or sufficiency of various conditions, proposed in the literature, for the existence of an organism. These include genetic homogeneity, sexual reproduction, bottlenecks, germ-soma separation, policing mechanisms, spatial boundaries, and immune response (Clarke Reference Clarke2013). Her position is that “the actual properties picked out by the classical views are red herrings, insofar as we are interested in finding general criteria for counting organisms” (Clarke Reference Clarke2012, 342).
Of course, Clarke does not deny that the proposed criteria are relevant to the issue of organismality. For her, though, each is one of a plurality of mechanisms that act, in various but not all circumstances, as realizers of one of two functional roles. It is these roles that, in conjunction, allow natural selection to act on the organism. The two functions are suppression of intraorganism selection via fitness-aligning policing mechanisms and “demarcation” mechanisms that promote interorganism selection via fitness variation. Taken together, instantiation of both functional roles is necessary and jointly sufficient for something to be a potential target of evolution by natural selection. I now examine the two functions in relation to pregnancy.
3.1. Policing
During pregnancy, two paradigmatic policing mechanisms are present with respect to both gestator and fetus, and hence one of Clarke’s requirements is met. Gestator and fetus are on opposite sides of a single-cell reproductive bottleneck, and both exhibit germ-soma division of reproductive labor. Taking these briefly in turn, mammalian reproduction proceeds via fusing of gametes to produce a single cell with a full complement of chromosomes, and all subsequent fetal cells are derived from repeated mitotic or meiotic divisions tracing back to that first bottleneck cell. Germ-soma sequestration in mammals (and many other organisms) refers to the existence of specialist germ cells that are the only cells capable of passing their genetic material to the next generation. Other cells, reproducing only via mitosis, are somatic cells, often referred to as evolutionary dead ends.Footnote 6
Clarke and others (Godfrey-Smith Reference Godfrey-Smith2009) demonstrate that single-celled bottlenecks are not necessary for organismality,Footnote 7 but Clarke’s account requires only that a sufficient combination of realizers be present. Bottlenecks can be single or multicellular, and the wider the bottleneck, the less it acts as a policing mechanism. In contrast, the single-cell bottleneck that begins a mammalian pregnancy is a highly, although not fully, effective mechanism for aligning fitnesses. In the absence of other effects, the bottleneck guarantees genetic homogeneity of cells, which eliminates fitness variation. Nevertheless, genetic mutations in somatic cells cause heterogeneity within mammals, and this is where germ-soma differentiation plays a role. The mechanism of germ-soma sequestration acts to damp down the potential for competition between genetically heterogeneous cells. The fact that somatic cells have zero fitness prevents them from participating as individuals in a process of natural selection.Footnote 8
I conclude that a single-cell bottleneck and germ-soma sequestration jointly act as powerful realizers of the policing function that aligns fitnesses of cells both within the gestator and within the fetus. These realizers are not ones Kingma (Reference Kingma2019) considers when she draws on theories of organismality to examine pregnancy.
3.2. Demarcation
I now move to Clarke’s “demarcation” mechanisms. These promote interorganism fitness differences within a population—variation that leads, when there is heritability of fitness, to evolution by natural selection. If mammals reproduced asexually, it could be questionable whether a demarcation mechanism exists. Taking Janzen’s (Reference Janzen1977) classic example, for dandelions and other asexual plants, what might be viewed as reproduction can be interpreted as growth due to the lack of genetic variation. This, in Clarke’s terms, prevents “demarcating” fitness differences. The same can be said for aphids undergoing reproduction by parthenogenesis, although Clarke argues that in this case other mechanisms are present that produce interorganism fitness variation (Reference Clarke2013, 428). The importance of sexual reproduction, and the recombination that goes with it, is decisively illustrated by the fact that its evolution is postulated to have occurred for precisely that reason. The evolution of sex is not easily explained, and a prevalent solution is that it promotes interorganism variation (Otto Reference Otto2008).
As with bottlenecking and germ-soma sequestration, Kingma’s examination of theories of organismality during pregnancy does not focus on sexual recombination as an individuating mechanism. The gestator and fetus are on either side of a variation-producing recombination event; hence, the second of Clarke’s two jointly sufficient roles is fulfilled.
That there are distinct fitness values for gestator and fetus is, of course, complicated by the fact that the expected fetal fitness appears as an element of the gestator’s fitness.Footnote 9 Nevertheless, gestator fitness is not identical to fetal fitness. For instance, the gestator could have large expected lifetime fitness (e.g., if she is highly fertile), but the fetus, a small lifetime expected fitness (if it possesses a trait that makes it unlikely to be carried for a full term of pregnancy). Similarly, the gestator could have low expected fitness (if she has a low probability of conceiving), while the fetus has a high expected fitness (where it does not inherit the low-fertility trait).
With both policing and demarcation mechanisms present, I conclude that, according to Clarke’s evolutionary view, the gestator and fetus constitute two organisms. Note that Clarke is willing to allow that organismality can be a matter of degree, rather than an all-or-nothing matter. This allowance is made for an account of the evolution of novel levels of individuation (Clarke Reference Clarke2013, 432). For instance, bottlenecks come in degrees, but in mammals the policing and demarcation mechanisms are not the marginal ones found in problem cases such as acacia groves and other plants made up of clonal ramets produced by vegetative growth via “runners” (Godfrey-Smith Reference Godfrey-Smith2009, 71); they are paradigmatic examples of the instantiation of Clarke’s two functional roles. If (evolutionary) organismality is a matter of degree, both gestator and fetus possess it to a high degree, possibly the highest degree. Having established that an evolutionary account of organismality yields the verdict that there are two during pregnancy, I move on to why a physiological account gives the verdict one.
4. Pregnancy and Immunological Organisms
Pradeu (Reference Pradeu2012) develops a detailed account of organismality based on physiology. Some of the criteria Clarke attacks for failing to be necessary or sufficient for organismality fall under the umbrella of physiology, but Pradeu’s account is one that cites immunological tolerance as the single defining physiological feature of organisms.Footnote 10 In fact, his account can be seen as challenging Clarke’s, on the grounds that it sees no need to move to a functional definition of organismality because there is, contrary to Clarke, a single mechanism underpinning individuation of all organisms.Footnote 11
Pradeu’s characterization is that organisms are the discrete, cohesive physiological biological entities whose boundaries and cohesiveness are marked out by what is immunologically tolerated. As he recently summarized, “any entity which interacts regularly with the immune system and is not eliminated by it is part of the physiological individual” (Pradeu Reference Pradeu2016b, 805).
These entities are very heterogeneous and include, for instance, tolerated microbial flora. Such heterogeneity is a central part of Pradeu’s theory and his rejection of the self-nonself theory of immunology. That theory has been very influential and interprets immune responses as differentiating between the organism’s own cells and foreign invaders (Pradeu Reference Pradeu2012, chap. 2). Pradeu supports his claim that this is a mischaracterization with the case of gut flora, along with the fact that some “self” cells such as cancerous ones can be rejected. Having established his view, I now consider how it can be used to count organisms during pregnancy.Footnote 12
It is well established that the gestator displays a high degree of immunotolerance of the fetus and vice versa.Footnote 13 Mechanisms of tolerance begin within a few days of fertilization, within 24–48 hours of the beginning of implantation of the fetus in humans (when it is a blastocyst; Moore et al. Reference Moore, Vidhya Nandan Persaud and Torchia2016, 23).Footnote 14 Tolerance is required because the fetus carries, for instance, paternal major histocompatibility antigens (Murphy Reference Murphy2012, 62). Murphy characterizes the fetus’s status thus: “The foetus is thus tolerated for two main reasons: it occupies a site protected by a nonimmunogenic tissue barrier, and it promotes a local immuno-suppressive response in the gestator. Several sites in the body, such as the eye, have these characteristics. … They are usually called immunologically privileged sites” (663).
Note the way in which the fetus is categorized with other parts of the gestator. Also, note that tolerance of the fetus is not merely down to a barrier (as was theorized in the past) but is an active, bidirectional process of the action of immunological factors. These include HLA-G antigens, regulatory T cells, and TGF-β and IL-10 cytokines (Wegmann et al. Reference Wegmann, Lin, Guilbert and Mosmann1993; Murphy Reference Murphy2012, 662; Pradeu Reference Pradeu2012, 112). Furthermore, fetal cells can persist and are tolerated by the gestator’s system, even many years after birth and vice versa, phenomena known respectively as fetal and maternal microchimerism (Williams et al. Reference Williams, Zepf, Longtine, Anchan, Broadman, Missmer and Hornstein2009; Jeanty, Derderian, and MacKenzie Reference Jeanty, Derderian and MacKenzie2014; Boddy et al. Reference Boddy, Fortunato, Sayres and Aktipis2015).Footnote 15 Hence, under Pradeu’s approach, the gestator-fetus appears to be one immunological organism. I say “appears” because I now want to consider two potential objections to my conclusion that the immunological approach gives a one-organism answer. I examine these only with regard to whether Pradeu’s theory does in fact give that answer during pregnancy. Potentially, these objections could be developed to argue that there is something wrong with Pradeu’s account if it yields this verdict, but, some brief comments aside, that is an issue for future work, and I focus here on confirming the verdict itself.
One might object that the fetus is not part of the gestator organism because it does not obviously play a functioning role in gestator physiology. In response, Pradeu is very clear that under his immunological approach, contribution to the whole is not relevant: “If one adopts my immunological perspective on physiological individuality, then the physiological individual admits among its constituents the resident microbes with which its immune system interacts and that it tolerates—regardless of their functional contribution to the whole” (Reference Pradeu2016b, 810).
A second potential objection refers to maternal-fetal evolutionary conflict theory. Gestator “tolerance” of the fetus might be interpreted, at least partly, as manipulation of the immune system by the fetus (Haig Reference Haig1993; McGovern et al. Reference McGovern2017). The thought here is that if entity X is manipulating organism Y into accepting it, then that undermines the case for X being a genuine part of Y. However, Pradeu is explicit that analogous immunological manipulation by parasites does not disallow them from being part of the “host’s” immunological organism: “Counterintuitively, it tells us that entities such as gut bacteria, skin bacteria or long-term tolerated parasites are parts of the mouse” (Reference Pradeu2012, 254).Footnote 16
As he suggests, such a position appears to flout folk-biological ideas about organisms and is also a point of disagreement with evolutionary approaches to organismality. If one is unsympathetic to Pradeu’s account, because of its judgment on parasites, then pregnancy may be another reason to be suspicious of the immunological approach. Yet, the case of pregnancy might be taken to be less unintuitive than that of parasite tolerance and provide some support for this more radical feature of the approach. As mentioned above, I leave examination of such arguments for future work.
5. Conclusion
My answer to the organism counting question during placental pregnancy is that an evolutionary approach yields the answer two but an immunological approach yields the answer one. When it comes to evolution by natural selection, a single-cell bottleneck and germ-soma specialization align fitness both within gestator and within fetus, and sexual recombination acts as a demarcation mechanism yielding variation in fitness between the two. Hence, both can be targets of a process of natural selection. By contrast, application of Pradeu’s immunophysiological theory leads to an equally firm but different conclusion regarding pregnancy. The extensive systems of immunotolerance acting across the placenta from the beginning of pregnancy until birth mark out one immunological organism.
Thus, placental pregnancy is an up-to-now unexplored example of a problem of counting organisms. This result is important because it diversifies discussion of organismality in mammals from gut flora and contested holobionts. Unlike gut flora, pregnancy is at the macroscale and involves only one species (Bourrat and Griffiths Reference Bourrat and Griffiths2018). It is also important because it applies to homo sapiens, making its absence from previous examinations striking given that noticing the exotic phenomenon of pregnant mammals did not even require the invention of the microscope.
In addition to reinforcing the fact that counting mammals is not as straightforward as has often been thought, I have offered a fuller examination of pregnancy and organismality than previous work on the metaphysics of pregnancy. As explained in section 2, Kingma (Reference Kingma2019) uses literature on organisms and individuality to argue for her part-whole mereological thesis. In doing so, she neglects to discuss the mechanisms that are most important in individuating evolutionary organisms. She does not consider bottlenecking, germ-soma differentiation, or sexual recombination. She does briefly consider immunology, and my detailed application of Pradeu’s theory, couched as it is in terms of parts of immune systems, does appear to support Kingma’s part-whole claim.Footnote 17
On might ask, finally, which approach to organismality is most relevant to the metaphysics of pregnancy? The evolutionary, two-organism answer to the counting question may be unsurprising. Given that natural selection theory deals with lifetime expected fitness, it is, in a sense, an atemporal analysis. It is not that development within a lifetime is ignored by the theory but, when organisms change over time (including being pregnant or a fetus), fitness values are amalgamated according to life-history theory (Charlesworth Reference Charlesworth1994). Hence, the gory details of developmental or physiological matters are, in a sense, invisible to the evolutionary calculus. By contrast, application of Pradeu’s concern with immunophysiology is, unsurprisingly, focused on the specific biology of placental pregnancy. One might well take from this that the evolutionary approach is wholly unsuited to the fields of obstetrics, midwifery, and veterinary medicine, although the theory of maternal-fetal conflict speaks against such a hasty, strong conclusion (Haig Reference Haig1993, Reference Haig2010). Nevertheless, there is a stark contrast between obstetric concerns and, for instance, theoretical biological work on evolutionary transitions of individuality, which seems wholly suited to Clarke’s approach. There, physiological change is to be explained via evolutionary theory, whereas the most pressing concern for vets, doctors, and midwives with pregnant patients is the physiological entwinedness in front of them.
